Risk and Return

in Roman Egypt
Paul Kelly
Email:
[email protected]
Agenda
• Introduction
• Underlying Research Questions
• Methodology
• Risk and Return Hierarchy
• Market Characteristics
• Overall Conclusions
 Introduction
• Papyri from Roman Egypt
• c. 20,000 documentary sources from AD 1 to 350, contracts,
official documents, receipts, letters ….
• Generally poorly attributed from an archaeological perspective
but found in variety of contexts:
• Houses
• Rubbish dumps
• Used as stuffing for mummified sacred crocodiles…
• 4,367 financial transactions involving 9,700+ individuals
• These are the records of ordinary people who worked the land
or lived in small cities – Note, no conclusions made about
Alexandria.

Map after Baines and Malek: 1980

Example Document: SB VIII 9830 – Lease of Land dated to AD 84-96

• [ἔτους ̣ ̣ Αὐτοκράτορος Καίσαρος

Δο]μιτιανοῦ Σεβαστοῦ Γερμανικοῦ
[Μεχεὶρ α ἐν Τεπτύνει τῆς Πολέμων(?)]ος
μερίδος τοῦ Ἀρσινοείτου νομοῦ. ὁμο-
[λογοῦσι Θέων Πάτρωνος ὡς ἐτῶν ̣ ̣ ο]ὐλὴ
ἀνκῶνι ἀριστερῶι κα̣ὶ Παποντῶς
[τοῦ δεῖνα ὡς ἐτῶν μϛ οὐλὴ μετώ]π̣ωι ἐξ
ἀριστερᾶς καὶ Σανσνεὺς Παπον-
5[τῶτος ὡς ἐτῶν ̣ ̣ οὐλη(?) ἀριστ]ερῶι καὶ
Ἀφροδίσιος Ἀφροδισίου, τοῦ Ἡρα-
[κλ -ca.?- ὡς ἐτῶν ̣ ̣ οὐλὴ δακτ]ύ̣[λω]ι μέ̣σ̣ωι̣ ̣
χειρὸς ἀριστερᾶς καὶ Ἀπολ-
[λώνιος(?) Ἀπολλωνίου ὡς ἐτῶν] Traces οὐλὴ
μετώπωι μέσωι καὶ Ἡρακλῆς
[Πτολεμαίου ὡς ἐτῶν ̣ ̣ οὐλὴ] μετώπωι
μέσωι καὶ Ἀπολλώνιος Ἀπολλωνί-
[ου ὡς ἐτῶν ̣ ̣ οὐλὴ δακτύλ]ωι πρώτωι
χειρὸς ἀριστερᾶς οἱ ἑπτὰ
10[γεωργοῦντες δημοσίων ἐδαφῶν…..

Photo: Photographic Archive of Papyri in the Cairo Museum

Underlying Research Questions
• What were
• The natures of the financial markets of Roman Egypt?
• Were they driven by reciprocity/redistribution….following
Weber/Polanyi?
• Or financial rationality…following Neoclassical economists?
• Financial rationality – use of reason to make comparative
decision to achieve financial goals having due regard to risk
• Need to understand the risks and returns of different
investment types
• Less about institutions, more about behaviours
• Less NIE more Behavioural Finance…
Methodology
• Models – ‘Some of the greatest fun I have had in Roman history has
been in constructing models…models allow us to perceive the
structures or repeated patterns which lie behind the superficial flow
of individual actors and events…models allow us to construct whole
pictures’ – Hopkins.
• Roman Egypt has the necessary background data to allow plausible
modelling, but given jump in prices c. AD 274 and high inflation
thereafter can generally only use data to AD 270.

6
Databases of Financial Transactions
• 6 inter-connected databases Transactions within Databases

• Information on terms, parties, 157

sizes etc gathered 482 88

• 4,367 transactions in total Land Sales

1273 368 Vacant Lots
House Sales
Loans and Deposits
Land Leases
House Lease
1999
Risk and Return Hierarchy
Options for investors – modelling
• Direct farming of land
• Leasing of land – 105 individual leases
• Housing for rent
• Secured cash loans – 67 individual loans where interest rate and
capital known
• Unsecured cash loans – 66 individual loans where interest rate and
capital known
• Commodity loans - 27 individual loans where interest rate known
• Trade and manufacture
Risk and Return Hierarchy
How to calculate investment return achieved?
• Analyse individual transactions to determine return which is

r = Profit divided by Capital employed

Profit = Income less frictional costs less taxation less depreciation/

maintenance

• Previous attempts, e.g. Carrié, have been based on single number or

range estimates.
Allowing for risk
• Assessing return without risk is misleading.
• Secured loan returns still have some residual risk of debtor default, in which case lender could
seize assets, which might even be of higher value than debt so make profit after costs
• Return from unsecured loans or land rental has however, considerable risks:
• Harvest failure or other risk leading to default or reduced rent
• Cash value of cereals received as rent or repayment will fluctuate
• And risks are inter-related – in good years when rent received in full, prices probably low, since
best prices available in years when harvest fails
• Requires a stochastic approach to model risks with series of Monte-Carlo simulations
• Value of cereals is inversely correlated with success of Nile flood
• Since emphasis on capital values not price – have not allowed for land price volatility
• Make 1,000 simulations of financial return for each transaction to establish range of possible
outcomes for this transaction
• Combine all simulations to arrive at likely distribution of returns
Risk and Return results Stochastic Return Distributions
80%
Return p.a. Median Average
70%

60%
Mortgages 8% 7%

Unsecured
Proportion of Cases

50%

Cash Loans 10% 0%

40%

30% Loans in kind 29% 24%

20%
Oxyrhynchite
10%
Land leases 7% 7%

0%
-1 -100% to - -5% to 0% 0% to 5% 5% to 10% 10% to 15% 15% to 20% 20% to 25% 25% to 30% 30% to 35% Over 35%
5%
Annual Return

Mortgages Unsecured Cash Loans Loans in kind Oxyrhynchite land leases

 Risk and Return Hierarchy
70.0%
Roman loans in
60.0% kind

50.0% Roman unsecured

cash loans
40.0%
Risk

30.0% US high yield bonds

US high grade US equities
20.0% bonds

10.0% US cash Roman land Roman mortgages

0.0%
0.0% 5.0% 10.0% 15.0% 20.0% 25.0% 30.0%
Return
Conclusions – Social Implications
• Earning 7% by leasing land or secured lending when inflation is
running at around 1% per annum, means doubling the real value of
your investments every 12 years.
• In ‘Piketty’ terms g, economic growth = inflation plus real growth.
Real growth in agrarian societies is generally below 1% per annum, so
g unlikely to exceed 2%.
• If r=7%, then “r>g”, the central precondition for concentration of
wealth and increased social inequality existed for most of the 270
years from AD 1.
Market Characteristics – Some remarks
• Seasonality – land leases and loan repayment dates highly seasonal –
leases concluded before sowing, loans repayable after harvest.
• Low intra-family transactions, with exception of house sales.
• Most transactions conducted by people in same location, ‘localisation
ratio’ < 1, with exception of land leases where often landlord is city-
based, tenant village-based.
• Markets were in some respects ‘limited access’ institutions – women
could not act as land tenants, literacy and geographic barriers to
accessing property records.
Market Characteristics – Questions
• Did all the key pre-requisites (sufficient information and competition,
low transaction costs) for a market driven by financial rationality
exist?
• Yes for land and credit, no for housing
• Were the financial outcomes of the markets rational?
• Yes for secured credit and land markets
• No for unsecured credit
• Do other indicators of financial rationality (e.g. marginal pricing,
contract structure, unrelated parties) exist?
• Yes for secured credit and land markets
Conclusions –
Nature of the Markets
• Cannot talk of a ‘single’ Roman market economy or bazaar, but
sectors and types of transaction which respond to different drivers.
• No evidence for financial rationality driving housing market.
• Outcome of unsecured lending is not financially rational so must be
driven by other features such as reciprocity.
• But pre-requisites exist, and evidence for, ‘exchange value’-driven
markets in land and secured lending given logical risk/return
hierarchy and other features.
Contact details and further reading
• Email: [email protected]
• See also Academia.edu site https://kcl.academia.edu/paulkelly
• This work derives from my recently completed doctorate at KCL – see
https://kclpure.kcl.ac.uk/portal/en/theses/search.html